A new epileptic seizure detection method based on fusion feature of weighted complex network

Abstract

Epileptic seizure detection plays an important role in the diagnosis of epilepsy. High-performance automatic detection method of epilepsy can reduce the workload of medical workers clinically, which has important clinical research significance. In this paper, we propose a new epileptic seizure detection method based on the statistic fusion feature of complex network. Firstly, we convert electroencephalogram (EEG) signals into complex network using horizontal visibility graph and weighted horizontal visibility graph. Then, we extract the average degree square and average weighted degree of complex network. Finally, the weighted sum of this two features is calculated as a single dimensional feature to classify the epileptic EEG signals. Experimental results show that the classification accuracy based on the feature fusion is up to 99%. It indicates that the classification accuracy of the single feature based on feature fusion is very high and the proposed method is effective to classify EEG signals.